Perceived weight-related stigma, loneliness, and mental wellbeing during COVID-19 in people with obesity: A cross-sectional study from ten European countries

Obesity is recognised as a chronic, complex disease by multiple organisations (e.g., AMA, WHO, EASO). Despite multifactorial causes (genetics, environment, gene–lifestyle interactions), societal beliefs often attribute obesity to individual behaviour, fostering weight-related stigma. Such stigma is linked to poorer mental wellbeing, including distress, social isolation, and avoidance of supportive activities. During COVID-19, people with obesity may have experienced worsened mental wellbeing due to infection fears, shielding and isolation, reduced service access, and extensive media coverage linking obesity to severe COVID-19. Media portrayals have been criticised for stigmatising language and blame, potentially exacerbating internalised stigma, shame, and psychological harm. Given immediate harms from stigma, mitigating actions are needed; loneliness may influence the impact of stigma on mental wellbeing, especially as pandemic restrictions heightened loneliness. The study investigated associations between perceived changes in representation of people with obesity in television and social media (as indicators of perceived weight-related stigma) and mental wellbeing during COVID-19 across 10 European countries. The authors hypothesised that poorer mental wellbeing would be associated with more negative perceived representation and that this association would be stronger among individuals with higher loneliness.

Design and data source: Cross-sectional online survey (EURopean Obesity Patient pANdemic Survey, EUROPEANS) conducted from September 3 to December 19, 2020, using Qualtrics. Recruitment via Qualtrics panels, recontact of opt-in participants, and patient-led obesity advocacy organisations (e.g., ECPO). Inclusion criteria: adults ≥18 years, self-reported BMI ≥30 kg/m², resident of England, Greece, Germany, France, Sweden, Israel, Portugal, Italy, Denmark, or Spain. Age-stratified quotas (18–30, 30–50, 50+). Respondents provided electronic informed consent. Ethics approval: University of Liverpool Research Ethics Committee (ref 7932). Data quality checks excluded speeding, patterned responses, poor verbatims, duplicates/bots, and incomplete datasets. Measures: - Demographics: gender, ethnicity, BMI (kg/m²), shielding status, educational attainment (as SES proxy). - Perceived change in representation of people with obesity in media: Single study-specific items asking whether representation during self-defined peak of COVID-19 exposure compared to pre-pandemic was more positive, more negative, or no change for (i) television and (ii) social media. Used as indicators of perceived experiences of weight-related stigma. - Mental wellbeing: PHQ-4 for depression and anxiety symptoms over past 2 weeks (two items each); higher scores indicate greater symptom burden; not diagnostic. Reliability in current sample: omega = 0.91; subscales Spearman-Brown: anxiety 0.87, depression 0.82. WHO-5 for subjective psychological wellbeing over past 2 weeks; higher scores indicate better wellbeing; omega = 0.93. - Loneliness: De Jong Gierveld (DJG) 6-item short scale assessing overall, emotional, and social loneliness over past 2 weeks; higher scores indicate greater loneliness; omega = 0.87. Statistical analysis: Analyses conducted in R 4.0.5. Linear mixed-effects models used with outcomes: total distress (PHQ-4 total), depression (PHQ-4 depression subscale), anxiety (PHQ-4 anxiety subscale), and wellbeing (WHO-5). Explanatory variables: perceived change in representation on television and social media (categorical: negative/none/positive; reference = no change). Covariates: age (years), gender (male vs female), BMI (kg/m²), shielding (yes/no), education (SES), ethnicity (White vs BAME), and total loneliness (DJG total). Random intercept for country significantly improved fit (χ²(1) > 19.88, p < 0.001; AIC differences > 17). Model fit comparison: control models (covariates only) vs full models (adding perceived representation variables) showed better fit for full models (AIC reductions; see Table 1). Residual diagnostics indicated normality; multicollinearity minimal (VIFs < 2.89). Robust standard errors were used in sensitivity analyses for heteroscedasticity in total distress models; results unchanged. Moderation: Interaction terms between mean-centred loneliness and dummy-coded more negative and more positive representation for both media types were tested in separate models for depression, anxiety, and wellbeing. Sample size and characteristics: N = 2882 adults with obesity; 56% female; mean age 40.6 ± 15.1 years; mean BMI 35.4 ± 5.9 kg/m²; ethnicity 92% White/White-mix. Country samples ranged from 276 to 297 participants.

- Sample: N = 2882; 56% female; mean age 40.6 years; mean BMI 35.4 kg/m²; 92% White/White-mix. - Model fit: Adding perceived media representation improved fit over control models for all outcomes (e.g., total distress AIC 14421 vs 14372; χ² diff 56.05, p < 0.001). - Television representation: Compared to no change, perceiving more negative TV representation was not significantly associated with depression (B = 0.09, 95% CI: -0.13, 0.31), anxiety (B = 0.19, 95% CI: -0.04, 0.41), wellbeing (B = 0.99, 95% CI: -1.79, 3.78), or total distress (B = 0.27, 95% CI: -0.13, 0.68). Perceiving more positive TV representation was associated with higher anxiety (B = 0.26, 95% CI: 0.06, 0.46), higher wellbeing (B = 4.52, 95% CI: 2.01, 7.02), and higher total distress (B = 0.45, 95% CI: 0.09, 0.81); not significantly associated with depression (B = 0.19, 95% CI: -0.01, 0.39; p = 0.054). - Social media representation: Compared to no change, perceiving more negative social media representation was associated with higher depression (B = 0.50, 95% CI: 0.29, 0.71), higher anxiety (B = 0.31, 95% CI: 0.10, 0.52), lower wellbeing (B = -3.52, 95% CI: -6.19, -0.86), and higher total distress (B = 0.80, 95% CI: 0.42, 1.19). Perceiving more positive social media representation showed no significant associations with depression, anxiety, wellbeing, or total distress. - Moderation by loneliness: Minimal evidence of moderation; only the interaction of more positive social media representation × lower loneliness predicting higher depression was significant (interaction B = -0.13, SE 0.06, p = 0.024), with very small variance explained (R² change ≈ 0.003). No significant moderation effects for anxiety or wellbeing. - Conditional R² for full models: ~0.25–0.30 across outcomes.

The study examined how perceived changes in the representation of people with obesity in media during COVID-19 related to mental wellbeing among adults with obesity across 10 European countries. Negative representation on social media was consistently associated with worse mental health (higher distress, anxiety, depression) and lower wellbeing, aligning with prior evidence that weight-related stigma harms psychological wellbeing. In contrast, negative representation on television showed no significant associations, possibly reflecting differences in how individuals engage with media: television content may primarily convey structural stigma with limited interaction, whereas social media can amplify structural, interpersonal, and intrapersonal stigma through interactive features and public commentary. Additionally, social media is comparatively less regulated than television, potentially allowing more overt stigmatising content. Interestingly, perceiving more positive representation on television was associated with both higher wellbeing and higher distress/anxiety. This mixed pattern may reflect simultaneous effects during the pandemic: increased reporting on obesity as a COVID-19 risk factor potentially elevating anxiety and distress, alongside coverage of supportive policies and strategies enhancing perceived support and wellbeing. Loneliness was hypothesised to exacerbate the impact of stigma on mental wellbeing; however, moderation analyses provided limited and inconsistent support, with negligible variance explained. Findings emphasise the importance of media accountability to avoid stigmatising portrayals and to promote respectful, accurate representations of people with obesity, which may help protect mental wellbeing.

Perceiving more negative representation of people with obesity on social media during the COVID-19 pandemic was associated with higher total distress and depression, and lower wellbeing among adults living with obesity. Perceiving more positive representation on television was associated with both greater wellbeing and higher distress/anxiety. Loneliness showed minimal moderating influence on these relationships. The authors encourage greater media accountability and the use of non-stigmatising portrayals to reduce harm to mental wellbeing among people with obesity. Future research should use longitudinal and mixed-methods designs to clarify causal pathways and explore potential moderators.

- Cross-sectional, retrospective, and subjective assessment of perceived changes in media representation limits causal inference and may introduce recall bias. - The timing of “peak” pandemic exposure was self-defined and varied within/between countries. - The measure of perceived media representation was study-specific (no COVID-19-specific validated instrument). - Potential reverse causality: poorer mental wellbeing may bias perception and interpretation of media content. - SES was approximated by education; more comprehensive SES measures (e.g., area-level deprivation) were not available. - Predominantly White sample (92%) limits generalisability to more diverse populations. - Although country-level differences were modelled via random intercepts, unmeasured confounding remains possible.